stpm (version 1.3.2)

spm_projection: A data projection with previously estimated or user-defined parameters. Projections are constructed for a cohort with fixed or normally distributed initial covariates.

Description

A data projection with previously estimated or user-defined parameters. Projections are constructed for a cohort with fixed or normally distributed initial covariates.

Usage

spm_projection(x, N = 100, ystart = 80, model = "discrete", tstart = 30, tend = 105, dt = 1, sd0 = 1, nobs = NULL, gomp = FALSE)

Arguments

x

A list of parameters from output of the spm(...) function.

N

A number of individuals to simulate, N=100 by default.

ystart

A vector of starting values of covariates (variables), ystart=80 by default.

model

A model type. Choices are: "discrete", "continuous" or "time-dependent".

tstart

Start time (age), default=30. Can be an interval: c(a, b) - in this case, the starting time is sumulated via runif(1, a, b).

tend

End time (age), default=105.

dt

A time interval between observations, dt=1 by default.

sd0

A standard deviation value for simulation of the next value of variable. sd0=1 by default.

nobs

A number of observations (lines) for i-th individual.

gomp

A flag (FALSE by default). When it is set, then time-dependent exponential form of mu0 and Q are used: mu0 = mu0*exp(theta*t), Q = Q*exp(theta*t). Only for continous-time SPM.

Value

An object of 'spm.projection' class with two elements. (1) A simulated data set. (2) A summary statistics which includes (i) age-specific means of state variables and (ii) Survival probabilities.

References

Yashin, A. et al (2007), Stochastic model for analysis of longitudinal data on aging and mortality. Mathematical Biosciences, 208(2), 538-551.

Akushevich I., Kulminski A. and Manton K. (2005). Life tables with covariates: Dynamic model for Nonlinear Analysis of Longitudinal Data. Mathematical Popu-lation Studies, 12(2), pp.: 51-80. .

Yashin, A. et al (2007), Health decline, aging and mortality: how are they related? Biogerontology, 8(3), 291-302..

Examples

## Not run:  
# library(stpm)
# # Setting up the model
# model.par <- list()
# model.par$a <- matrix(c(-0.05, 1e-3, 2e-3, -0.05), nrow=2, ncol=2, byrow=TRUE)
# model.par$f1 <- matrix(c(90, 35), nrow=1, ncol=2)
# model.par$Q <- matrix(c(1e-8, 1e-9, 1e-9, 1e-8), nrow=2, ncol=2, byrow=TRUE)
# model.par$f <- matrix(c(80, 27), nrow=1, ncol=2)
# model.par$b <- matrix(c(6, 2), nrow=2, ncol=2)
# model.par$mu0 <- 1e-6
# model.par$theta <- 0.09
# # Projection
# # Discrete-time model
# data.proj.discrete <- spm_projection(model.par, N=5000, ystart=c(80, 27))
# plot(data.proj.discrete$stat$srv.prob)
# # Continuous-time model
# data.proj.continuous <- spm_projection(model.par, N=5000, ystart=c(80, 27), model="continuous")
# plot(data.proj.continuous$stat$srv.prob)
# # Time-dependent model
# model.par <- list(at="-0.05", f1t="80", Qt="2e-5", ft="80", bt="5", mu0t="1e-3*exp(0.08*t)")
# data.proj.time_dependent <- spm_projection(model.par, N=500, ystart=80, model="time-dependent")
# plot(data.proj.time_dependent$stat$srv.prob)
# ## End(Not run)